Allow any XML mimetype in Connection

Fixes #2059
jhy · Nov 23, 2023 · 58521a4 · 58521a4
1 parent bc79810
commit 58521a4
Show file tree

Hide file tree

Showing 4 changed files with 218 additions and 6 deletions.
diff --git a/CHANGES b/CHANGES
@@ -37,6 +37,10 @@ Release 1.17.1 [PENDING]
     which caused that pinning.
     <https://github.com/jhy/jsoup/issues/2054>
 
+  * Improvement: in Jsoup.Connect, allow any XML mimetype as a supported mimetype. Was previously limited to
+    {application|text}/xml. This enables for e.g. fetching SVGs with a image/svg+xml mimetype, without having to
+    disable mimetype validation.
+
   * Bugfix: when outputting with XML syntax, HTML elements that were parsed as data nodes (<script> and <style>) should
     be emitted as CDATA nodes, so that they can be parsed correctly by an XML parser.
     <https://github.com/jhy/jsoup/pull/1720>

diff --git a/src/main/java/org/jsoup/helper/HttpConnection.java b/src/main/java/org/jsoup/helper/HttpConnection.java
@@ -798,9 +798,9 @@ public static class Response extends HttpConnection.Base<Connection.Response> im
         private final HttpConnection.Request req;
 
         /*
-         * Matches XML content types (like text/xml, application/xhtml+xml;charset=UTF8, etc)
+         * Matches XML content types (like text/xml, image/svg+xml, application/xhtml+xml;charset=UTF8, etc)
          */
-        private static final Pattern xmlContentTypeRxp = Pattern.compile("(application|text)/\\w*\\+?xml.*");
+        private static final Pattern xmlContentTypeRxp = Pattern.compile("(\\w+)/\\w*\\+?xml.*");
 
         /**
          <b>Internal only! </b>Creates a dummy HttpConnection.Response, useful for testing. All actual responses
@@ -885,7 +885,7 @@ else if (methodHasBody)
                         && !contentType.startsWith("text/")
                         && !xmlContentTypeRxp.matcher(contentType).matches()
                         )
-                    throw new UnsupportedMimeTypeException("Unhandled content type. Must be text/*, application/xml, or application/*+xml",
+                    throw new UnsupportedMimeTypeException("Unhandled content type. Must be text/*, */xml, or */*+xml",
                             contentType, req.url().toString());
 
                 // switch to the XML parser if content type is xml and not parser not explicitly set

diff --git a/src/test/java/org/jsoup/integration/ConnectTest.java b/src/test/java/org/jsoup/integration/ConnectTest.java
@@ -4,13 +4,16 @@
 import org.jsoup.HttpStatusException;
 import org.jsoup.Jsoup;
 import org.jsoup.Connection.Method;
+import org.jsoup.UnsupportedMimeTypeException;
 import org.jsoup.helper.DataUtil;
 import org.jsoup.helper.W3CDom;
 import org.jsoup.integration.servlets.*;
 import org.jsoup.internal.StringUtil;
 import org.jsoup.nodes.Document;
 import org.jsoup.nodes.Element;
 import org.jsoup.nodes.FormElement;
+import org.jsoup.nodes.Node;
+import org.jsoup.nodes.XmlDeclaration;
 import org.jsoup.parser.HtmlTreeBuilder;
 import org.jsoup.parser.Parser;
 import org.jsoup.parser.XmlTreeBuilder;
@@ -509,17 +512,17 @@ public void handlesWrongContentLengthDuringBufferedRead() throws IOException {
     }
 
     @Test
-    public void testBinaryContentTypeThrowsException() {
+    public void testBinaryContentTypeThrowsException() throws IOException {
         Connection con = Jsoup.connect(FileServlet.urlTo("/htmltests/thumb.jpg"));
         con.data(FileServlet.ContentTypeParam, "image/jpeg");
 
         boolean threw = false;
         try {
             con.execute();
             Document doc = con.response().parse();
-        } catch (IOException e) {
+        } catch (UnsupportedMimeTypeException e) {
             threw = true;
-            assertEquals("Unhandled content type. Must be text/*, application/xml, or application/*+xml", e.getMessage());
+            assertEquals("Unhandled content type. Must be text/*, */xml, or */*+xml", e.getMessage());
         }
         assertTrue(threw);
     }
@@ -540,6 +543,26 @@ public void testBinaryContentTypeThrowsException() {
         assertEquals(Document.OutputSettings.Syntax.xml, doc.outputSettings().syntax());
     }
 
+    @Test public void imageXmlMimeType() throws IOException {
+        // test that we switch to XML, and that we support image/svg+xml
+        String mimetype = "image/svg+xml";
+
+        Connection con = Jsoup.connect(FileServlet.urlTo("/htmltests/osi-logo.svg"))
+            .data(FileServlet.ContentTypeParam, mimetype);
+        Document doc = con.get();
+
+        assertEquals(mimetype, con.response().contentType());
+        assertTrue(doc.parser().getTreeBuilder() instanceof XmlTreeBuilder);
+        assertEquals(Document.OutputSettings.Syntax.xml, doc.outputSettings().syntax());
+        Node firstChild = doc.firstChild();
+        XmlDeclaration decl = (XmlDeclaration) firstChild;
+        assertEquals("no", decl.attr("standalone"));
+        Element svg = doc.expectFirst("svg");
+        Element flowRoot = svg.expectFirst("flowRoot");
+        assertEquals("flowRoot", flowRoot.tagName());
+        assertEquals("preserve", flowRoot.attr("xml:space"));
+    }
+
     @Test
     public void canFetchBinaryAsBytes() throws IOException {
         Connection.Response res = Jsoup.connect(FileServlet.urlTo("/htmltests/thumb.jpg"))